Getting Started with SKVector

Overview

SKVector provides semantic vector search for AI memories using Qdrant. Find memories by meaning, not just keywords.

Installation


pip install "skmemory[skvector]"

Or standalone:


pip install skvector

Requirements

Quick Start


from skvector import VectorStore

vs = VectorStore(url="http://localhost:6333", collection="memories")

# Index a memory
vs.index(
    id="mem-001",
    text="Discussed the sovereign identity architecture with the team",
    metadata={"importance": 0.8, "emotions": ["excited"]}
)

# Semantic search
results = vs.search("identity protocol design", limit=5)
for result in results:
    print(f"{result.id}: {result.score:.2f} — {result.text[:80]}")

Integration with SKMemory

SKVector works as a search backend for SKMemory:


from skmemory import Memory

mem = Memory(vector_backend="qdrant", vector_url="http://localhost:6333")

# Memories are automatically vectorized and indexed
mem.snapshot(title="Protocol design", content="The challenge-response flow...")

# Semantic search finds by meaning
results = mem.search("authentication flow", mode="semantic")

Embedding Models

SKVector supports multiple embedding providers: